Analyzing User Behavior Data in a Mobile Tennis Game

The users of online multiplayer games generate vast amount of data that can be later used to fine-tune the game and optimize user experience. The present paper is dedicated to the analysis of user behavior in a mobile tennis game World of Tennis: Roaring '20s. It implements a pseudo-multiplayer game mode by introducing AI-controlled opponents trained on real user data and mimicking human behavior. We analyze user recordings to ascertain they exhibit a variety of consistent play styles that can be learned by the AI system to ensure challenging and lasting entertainment for the players.